Maurice Rabb wrote:
....
> BTW, I see that it was only Hans-Martin Mosner's experimental system that
> uses more than a single tag bit.  Hans, is your website temporarily down?
> I haven't been able to get on for the past few days?

It's not down, but our new Internet provider goofed up and so our name
disappeared from the domain name system. BTW, it's the german Telekom...
It's now an administrative problem, not a technical one, so it might
need some days to be sorted out.
In the meanwhile, you can always get at our server under its numeric IP
address:
www.heeg.de = 192.109.54.18

Regarding the subject (Scientific Computing):
I would consider immediate floats somewhat useless. Their precision is
too limited, and getting them into the FPU registers is just as costly
as reading a (single precision) float from an object's data. For
graphics, fixed point numbers make much more sense, and they should be
supported if we go to more tag bits.
Typical numerical algorithms deal with arrays of numerical data anyway,
so a FloatArray and a DoubleArray would be a sound basis for them.

Hans-Martin

I'll be on vacation for the next 3 weeks, so don't expect any mail from
me in the meantime...